The Selke finalists: a microstatic look

A couple of modern traditions are taking place for hockey watchers this week. On a daily basis the NHL is releasing the names of the "finalists" of its major trophies, the three players who received the most votes for the Hart, Norris, Vezina, Calder, and Lady Byng Trophies, among others. It's a bit of a gimmick, since there are no "finals"; the order of finish is already known, but won't be released until the awards gala at the end of the season. The process does serve to highlight the fine seasons of more than just the winners.

The other tradition is that of second-guessing the voters. Which players made the short list on the strength of their reputation rather than their actual play? Who's the "seen him good", or for that matter the "remember him good", darling of the media? Who's the up-and-comer who is getting touted for an award before he's quite there yet? Who's the deserving guy on the local team who yet again has been cruelly overlooked?

One award that has always prompted dark mutterings about the competency of the voters is the Frank J. Selke Trophy for the best defensive forward. Defensive efficiency is notoriously difficult to measure, so it has always seemed to be more of a reputation award than anything. While many deserving players have been bypassed over the years * coughJariKurricough * the list of Selke winners stands up to scrutiny pretty darn well, featuring as it does many great checkers and two-way players and only a couple "flavour of the month" duds like Steve Kasper and his minus-18.  

It seems to me that defensive play is one aspect of the game that we surely have a better handle on nowadays. There are many more statistical categories that capture a player's role within the game and within his team, from the NHL's own much-maligned RTSS numbers to some of the more reliable metrics gathered and disseminated by keen observers of the game like Gabe Desjardins and Vic Ferrari. We now have reliable data on where a player is deployed (ZoneStart), against whom (Quality of Competition), and how the play flows when he is out there (Corsi).

In 2009-10 the Selke finalists are an interesting group. Pavel Datsyuk, Ryan Kesler, and Jordan Staal all played on solid playoff-bound teams that finished 4th or 5th in their respective conferences with 101-103 points. Datsyuk and Staal have faced off in the last two Stanley Cup Finals, while Kesler's team reached the SCF as recently as 1994 ... oops, maybe one of these things is not like the others.  :)

All are centres, although in somewhat different roles within their teams; for discussion purposes let's consider Datsyuk a first-line centre, Kesler a second-liner, and Staal the prototypical third-line checking centre. After the jump we'll have a look at their relative performances across a broad spectrum of statistical metrics.

I have tried to group these stats in logical clumps, showing both raw totals and where the player ranks among the forwards with 40+ GP on his own team in each category. My two primary sources were NHL.com and behindthenet.ca; the stats from the former show the players ranking as a single number, while the latter shows that ranking against the number of eligible teammates (13 in each case; I removed Alexei Ponikarovsky from the Pittsburgh equation as he only played 16 games with the Pens). The PK stats at bottom show only those forwards meeting the dual standards of 40+ GP and 1:00+ SH TOI/G. I chose to omit powerplay statistics altogether as they have little bearing on defensive performance. Note also that I have shown in bold, those categories which I deem to be purely or primarily defensive in nature, and in italics those which measure two-way play. 

First of all, the similarities. All three are big-minute centres, ranking 1st, 2nd and 3rd respectively on their teams in TOI, in keeping with their primary roles on the 1st, 2nd and 3rd lines respectively. All three lead their squads in Shifts per game, which suggests extra defensive assignments for Kesler and Staal. All three had very similar offensive production at even strength, with the discrepancy in their overall totals due to Staal's lack of powerplay time and production. All three have very solid plus numbers in Relative Corsi despite facing competition and ZoneStarts that are well above team averages in all cases.

The case for Datsyuk: The Red Wings star is gunning for his third consecutive Selke and the numbers make a compelling argument. Once again Datsyuk led the loop in Takeways by a very wide margin, almost 60% ahead of Kesler who ranked second in the NHL. Of course such numbers must be taken with a grain of salt given the apparent lack of a league-wide standard, but it's hard to deny Datsyuk is a master of the art of puck-stealing. Particularly impressive is his ability to challenge for the puck without taking penalties; not only will that "gentlemanly and effective" combination help him win his fifth straight Lady Byng, it speaks directly to excellent defence that he doesn't often put his team at a manpower disadvantage. Datsyuk has the best GA/60 of any of the three candidates.

Contra: Datsyuk faces slightly lower QualComp than his rivals, and also plays with top-level teammates. His penalty-killing role is much less than the others, surely an important consideration for a defensive player. His great Relative Corsi in that situation has to be discounted significnatly due to small sample size; it's not that he can't kill penalties, it's just that he doesn't.

 
The case for Kesler: By raw numbers (although not team ranking) Kesler had the toughest ZoneStart of the three, with an O-zone Pct. below 45% compared to the others' ~50%. He was by far the leading shot blocker, and the most effective faceoff man overall, taking many more draws than Datsyuk and achieving a much better percentage than Staal. He has a great Tk:Gv ratio which can only partially be written off to home scorer bias. He faces the toughs playing with second-level teammates.

Contra: Kesler was a minus player at even strength, with a lousy ES GA of over 3 per 60,almost a goal per hour worse than his rivals and ranking a lowly 12th on his own club. This mirrors his poor numbers in on-ice Sv% and PDO#, which depending on how you feel about the dice roll of percentages, either explains away Kesler's poor numbers or is symptomatic of too many gold-star scoring chances against on his watch. Kesler also takes more penalties than his rivals, though he draws more as well.

The case for Staal: This guy is a very balanced player, ranking among the top 6 forwards on his team in every single category listed, and in the top 3 in a rank majority of those. He's a horse on the PK, ranking second in the NHL (behind only Jay McClement) in SH TOI . Like Kesler, he plays top level opposition with second level linemates; but in Jordan's case he manages to outscore by a healthy margin in the process.

Contra: Staal's faceoff percentage is still on the shady side of 50%; although he has significantly improved in this category every year he's been in the league - he was a Coglianoesque 37.1% as a rookie in '06-07 - this is one indicator that this guy is still a kid (just 21 this season).

My "vote": Of course I don't have one, that said I daresay I've already put more time into researching this than 95% of the PHWA members who do cast ballots. :p  If I was to throw out "seen him good" and simply rely on number crunching, one method might be to simply identify a basket of key categories and see how the guy ranks on his own team across the board. For argument's sake I would tend toward a disparate sixpack of metrics that includes QualComp, ZoneStart, SH TOI, GA On/60, Relative Corsi, and Tk:Gv ratio, with the first three identifying the player's role on his team and the last three crudely measuring his performance both with and without the puck. If one was to apply this method to this year's candidates, the surprise winner of the Selke Trophy would be Jordan Staal.

Those choices (and omissions) are by nature arbitrary. No doubt arguments can be raised to include other metrics, or to use a weighting system; I welcome discussion in the comments section. Being a lazy bastard at heart I have only done the detailed comp of the three finalists, so I also invite you to submit names of other worthy Selke candidates - Mike Richards? Martin Hanzal? Brandon Sutter? - who might have similar or even better statistical qualifications than the above three. Nonetheless, I am satisfied that the information at hand confirms that each of the three finalists has staked a solid claim to the trophy.